From 2c73e7dc297cc2669a5f47db00ae914347ab996e Mon Sep 17 00:00:00 2001 From: David Blodgett Date: Fri, 4 Aug 2017 16:43:57 -0500 Subject: [PATCH] Noticed a version type in vignette and a couple things in readme. --- README.md | 4 ++-- inst/doc/packageDiscrepencies.Rmd | 6 +++--- inst/doc/packageDiscrepencies.html | 6 +++--- vignettes/packageDiscrepencies.Rmd | 6 +++--- 4 files changed, 11 insertions(+), 11 deletions(-) diff --git a/README.md b/README.md index e3732d6..9dad3c5 100644 --- a/README.md +++ b/README.md @@ -29,9 +29,9 @@ Once you've restarted R, you can do: And you can update with: `update.packages()` -It is also possible to install without add GRAN to your rprofile, but automatic package updates will not be installed with `update.packages()`. +It is also possible to install without adding GRAN to your rprofile, but automatic package updates will not be installed with `update.packages()`. ``` -install.packages("smwrData", repos=c("https://owi.usgs.gov/R",getOption("repos"))) +install.packages("EflowStats", repos=c("https://owi.usgs.gov/R",getOption("repos"))) ``` More details are available at the [GRAN web page.](https://owi.usgs.gov/R/gran.html) diff --git a/inst/doc/packageDiscrepencies.Rmd b/inst/doc/packageDiscrepencies.Rmd index d365839..c67fd3b 100644 --- a/inst/doc/packageDiscrepencies.Rmd +++ b/inst/doc/packageDiscrepencies.Rmd @@ -9,7 +9,7 @@ vignette: > \usepackage[utf8]{inputenc} --- -This vignette summarizes discrepancies between the statistic calculations in the new (>v1.0) and the old (v1.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the [Closed issues](https://github.com/USGS-R/EflowStats/issues?q=is%3Aissue+is%3Aclosed) page of the EflowStats GitHub repository. +This vignette summarizes discrepancies between the statistic calculations in the new (>v5.0.0) and the old (v5.0.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the [Closed issues](https://github.com/USGS-R/EflowStats/issues?q=is%3Aissue+is%3Aclosed) page of the EflowStats GitHub repository. ## Issues in original HITHAT software @@ -17,7 +17,7 @@ The original HITHAT software contained two bugs in the code that affected a numb ## Old EflowStats package issues -There are a number of discrepancies between the new (v>1.0) and old (v<1.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>1.0) will be documented in the package _NEWS_ file. +There are a number of discrepancies between the new (v>5.0.0) and old (v<5.0.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>5.0.0) will be documented in the package _NEWS_ file. ### DH and DL statistics @@ -29,7 +29,7 @@ A second change affects the DH12-21 statistics. The original documentation in He Two bugs in the old EflowStats code affect some of the MA statistics. One minor issue caused incorrect rounding for the MA24-35 statistics and has been corrected in the new package. Another bug miscalculated the coefficient of variation and affected the MA3 statistic. This has been corrected in the new package. ### Magnificent Seven statistics -The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v1.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v1.0) package. +The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v5.0.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v5.0.0) package. ## General changes in the new EflowStats diff --git a/inst/doc/packageDiscrepencies.html b/inst/doc/packageDiscrepencies.html index dc9588a..0f259ea 100644 --- a/inst/doc/packageDiscrepencies.html +++ b/inst/doc/packageDiscrepencies.html @@ -35,14 +35,14 @@

04 August, 2017

-

This vignette summarizes discrepancies between the statistic calculations in the new (>v1.0) and the old (<v1.0) EflowStats packages. In general, the authors focused on ensuring that the statistics were calculated according to the definitions stated in Henriksen et al, 2006, rather than on reproducing output from previous package versions or the original HITHAT software. This process uncovered a number of issues with both the old EflowStats (>v1.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the Closed issues page of the EflowStats GitHub repository.

+

This vignette summarizes discrepancies between the statistic calculations in the new (>v5.0.0) and the old (<v5.0.0) EflowStats packages. In general, the authors focused on ensuring that the statistics were calculated according to the definitions stated in Henriksen et al, 2006, rather than on reproducing output from previous package versions or the original HITHAT software. This process uncovered a number of issues with both the old EflowStats (>v5.0.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the Closed issues page of the EflowStats GitHub repository.

Issues in original HITHAT software

The original HITHAT software contained two bugs in the code that affected a number of the statistics. One involved the calculation of summary monthly statistics where a single month statistic was returned rather than the period of record summary for that month. This bug affects many statistics that rely on monthly period of record summaries, such as the MA12-23 statistics. A similar bug was found with 7-day rolling mean calculations and influenced the ML17-18 statistics.

Old EflowStats package issues

-

There are a number of discrepancies between the new (v>1.0) and old (v<1.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>1.0) will be documented in the package NEWS file.

+

There are a number of discrepancies between the new (v>5.0.0) and old (v<5.0.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>5.0.0) will be documented in the package NEWS file.

DH and DL statistics

There are two changes to the method for calculatings means that affect many of the DH and DL statistics. First, rolling means in the old EflowStats package were calculated on a yearly basis. This makes little sense because rolling means are influenced when the windowing period approaches the beginning or end of the record. By splitting the period of record on year values, it compounds the “edge of record” issue by creating edges for every yearly time series rather than just at the beginning and end of the period of record. The bias can be substantial for longer window (30 and 90-day) rolling means. Thus, rolling means are calculated in the new EflowStats on the entire period of record before splitting the record into yearly time series for subsequent calculations.

@@ -54,7 +54,7 @@

MA statistics

Magnificent Seven statistics

-

The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v1.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v1.0) package.

+

The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v5.0.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v5.0.0) package.

diff --git a/vignettes/packageDiscrepencies.Rmd b/vignettes/packageDiscrepencies.Rmd index d365839..c67fd3b 100644 --- a/vignettes/packageDiscrepencies.Rmd +++ b/vignettes/packageDiscrepencies.Rmd @@ -9,7 +9,7 @@ vignette: > \usepackage[utf8]{inputenc} --- -This vignette summarizes discrepancies between the statistic calculations in the new (>v1.0) and the old (v1.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the [Closed issues](https://github.com/USGS-R/EflowStats/issues?q=is%3Aissue+is%3Aclosed) page of the EflowStats GitHub repository. +This vignette summarizes discrepancies between the statistic calculations in the new (>v5.0.0) and the old (v5.0.0) statistics and the original HITHAT software (Henriksen et al, 2006). A record of this process can be found in the [Closed issues](https://github.com/USGS-R/EflowStats/issues?q=is%3Aissue+is%3Aclosed) page of the EflowStats GitHub repository. ## Issues in original HITHAT software @@ -17,7 +17,7 @@ The original HITHAT software contained two bugs in the code that affected a numb ## Old EflowStats package issues -There are a number of discrepancies between the new (v>1.0) and old (v<1.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>1.0) will be documented in the package _NEWS_ file. +There are a number of discrepancies between the new (v>5.0.0) and old (v<5.0.0) EflowStats packages. A list of these by statistic is provided below, but any additional changes to the EflowStats statistics in subsequent versions (>5.0.0) will be documented in the package _NEWS_ file. ### DH and DL statistics @@ -29,7 +29,7 @@ A second change affects the DH12-21 statistics. The original documentation in He Two bugs in the old EflowStats code affect some of the MA statistics. One minor issue caused incorrect rounding for the MA24-35 statistics and has been corrected in the new package. Another bug miscalculated the coefficient of variation and affected the MA3 statistic. This has been corrected in the new package. ### Magnificent Seven statistics -The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v1.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v1.0) package. +The AR1 coefficient was miscalculated in the old EflowStats package. A bug in the old package caused some of the data to become disordered in time, which substantially influenced the calculation of the AR1 statistic. This has been fixed in the new package (>v5.0.0). Additionally, the phase calculation in the old EflowStats package was returned in radians, which was not very useful for interpretation. Thus, the value is now returned in units of days in the new (>v5.0.0) package. ## General changes in the new EflowStats